Ted Kane

Software Engineer | Biometric Researcher | Data Scientist

Hi, I'm Ted Kane, a software engineer experienced in the field of biometric recognition and machine learning.

Master of Electrical and Computer Engineering, Clarkson University (May 2024)

Bachelor of Software Engineering and Minor in Mathematics, Clarkson University (May 2022)

Biometrics
- Modalities I've worked with Face, Fingerprint, Iris, Voice, Gait, Keystroke, Mouse, and other behavioral modalities. For most of these, I've developed and trained my own deep learning models, and for some I've created end-to-end user systems.
- Biometric Template Protection (BTP): Systems designed to compare biometric samples while preserving the privacy of both samples. I've worked with Biometric Cryptosystems and briefly touched on Fully Homomorphic Encryption.
- Presentation Attack Detection (PAD): Systems designed to detect presentation attacks to biometric sensors such as face masks, pictures, and deepfakes. I've created PAI, evaluated commercial PAD systems, and developed my own liveness detection.

Machine Learning
- Problem Types I've worked with: Classification, Regression, Segmentation, Embedding, AutoEncoding, Clustering
- Data Types I've worked with: Images, 3D Positional Data, Audio, Financial, Sequential Behavioral
- Familiar Libraries: PyTorch, Tensorflow, Scikit-Learn, Flask, Pandas, Numpy, CuPY, Matplotlib

AntiCheats: Automating the detection of unfair advantages in videogames.
- Digital Forensics: Anticheat software designed to run on-device which detects suspicious executables and behavior. I've created my own tool for this (Kangaroo) in C#, which analyzes process memories, windows event history, and more to detect traces of failed cleanup.
- Server-Side AntiCheats: Software that runs on the server which detects unfair advantages based on information sent from the client. My own work in this involves developing a simulation based anticheat in Java